تلاش برای جستجوی کار مانیتورینگ: ارزیابی بر اساس یک طرح انقطاع رگرسیون

Since July 2004, the job search effort of long-term unemployed benefit claimants has been monitored in Belgium. We exploit the discontinuity in the treatment assignment at the age of 30 present in the first year of the reform to evaluate the effect of a notification sent at least eight months before job search is verified. Eight months after this notification and prior to the first monitoring interview, transitions to employment have increased by nearly nine percentage points compared to the counterfactual of no reform. Participation in training is raised, but not significantly, while withdrawals from the labor force are not affected.

The payment of unemployment benefits (UB) involves a trade-off between insurance and work incentives. Many economic researchers have studied how limiting the coverage of UB and the duration of benefit entitlement can restore work incentives (e.g. see Lalive et al., 2006). However, most UB schemes also provide work incentives by imposing job search requirements, amongst others, on benefit claimants. Boone et al. (2007) argue that such requirements, enforced by monitoring and sanctions, may deliver the right incentives by imposing lower costs than the aforementioned alternatives. This paper is about the impact of such a monitoring scheme introduced in the Belgian UB scheme in 2004. In many countries monitoring of job search effort is organized along relatively standardized procedures (OECD, 2007). It starts off with a notification (often at initial registration) by which the unemployed worker is informed about the search requirements and the proofs thereof to be delivered, about the timing of the evaluations of search effort, and about the associated sanctions in the case of non-compliance. On the prescribed dates, past job search effort is evaluated on the basis of transmitted paper proofs of job applications or in face-to-face interviews. If the outcome of the evaluation is negative, a sanction in the form of a temporary and partial reduction of unemployment benefits (UB) usually follows. Early studies1 found positive effects of monitoring on the job finding rate. However, since programs themselves often combined counseling with monitoring, they could not disentangle which of these components was responsible for such findings. A number of later contributions have succeeded in isolating the pure effects of monitoring. Klepinger et al. (1997) in the US and McVicar (2008) in Northern Ireland demonstrate that monitoring significantly increases transitions to employment.2 In contrast to this evidence, Ashenfelter et al. (2005) find that tighter search requirements in the US have insignificant effects on transitions to employment and Klepinger et al. (2002) report even slightly decreasing job finding rates. This is in line with the insignificant effect of job search monitoring reported by van den Berg and van der Klaauw (2006) for the Netherlands. Van den Berg and van der Klaauw (2006) argue that the result is caused by the substitution of formal by informal search, a phenomenon that would be especially relevant for well qualified workers on whom they focus in their study. Finally, Manning (2009) reports that too strict search requirements may lead UB recipients to stop claiming and to withdraw from the labor force. Petrongolo (2009) confirms this, demonstrating moreover that monitoring substantially decreases employment stability and annual earnings in the long‐term. In Belgium job search effort has only been monitored since 2004 and it targets only long-term unemployed workers, collecting UB for more than 13 months. Monitoring consists of face-to-face interviews in which caseworkers have a reasonable degree of discretion in the evaluation of the fulfillment of search requirements. The system is more lenient than in many other countries in that evaluations are much more spread out over time and the first negative evaluation does not lead to a monetary sanction. In addition, at the time of notification, job search requirements are not, as is usual, stated in terms of delivering proof of a minimum number of job applications, but it is rather vaguely stated that one needs to be searching for jobs on a “regular basis” and to collect written proofs of the search actions undertaken. By contrast, if imposed, sanctions are substantial. If one does not comply with the search actions stipulated at the first negative evaluation, benefits can be completely withdrawn at the subsequent interview: first temporarily during four months, but subsequently the entitlement to UB is completely halted in the case of recidivism. In this paper we evaluate the impact of the 2004 reform in Flanders (the Dutch speaking northern region of Belgium).3 The analysis is based on rich administrative data, which do not only allow the impact on the job finding rate to be identified, but also on exits to training and to a residual “out of the labor force” state.4 A regression discontinuity design (RDD), resulting from the gradual phasing in of the new monitoring scheme by age group, identifies these effects under weak assumptions (Lee and Lemieux, 2010). Between July 2004 and June 2005 the job search requirements were only imposed on benefit claimants younger than 30 years on July 1, 2004. In the subsequent years, the older age groups were gradually integrated. This study exploits the discontinuity in the treatment assignment at the age of 30. Since this discontinuity disappears after a year, we can only identify the effect of monitoring job search from the moment that the assessment period starts until eight months later, before the first monitoring of past job search effort takes place. We therefore cannot study the impact of the assessments at interviews and the imposed sanctions in the case of a negative evaluation. The outline of the article is as follows. In the next section we describe the institutional setting and the features of the new monitoring scheme. Section 3 describes the data. The estimation method is presented in Section 4. Section 5 reports the treatment effects on various outcomes and contains a number of validity checks. A final section contains the conclusion.

This paper studied the impact of an important reform in Belgium which introduced for the first time job search requirements into the UB scheme. More specifically, it evaluated the effect on various outcome variables (employment, participation in training and labor force withdrawal) of a notification announcing that job search effort will be retrospectively monitored at least eight months later. Eight months after this notification and prior to the first evaluation of job search effort the likelihood of a transition to employment is nearly 9 pp higher than in the absence of the reform. This finding is robust, but not precisely estimated. In view of the relatively lenient monitoring system in which interviews are conducted at a low frequency, the job search requirements are not clearly defined and a benefit sanction is not imposed in the case of non-compliance at the first interview, this high treatment effect may come as a surprise. Nevertheless, this does not mean that one should expect no effect of this notification. Based on a non-stationary job search model that takes the main features of the Belgian monitoring scheme into account, Cockx et al. (2011) find that an increase of 2 pp is consistent with the rational behavior of individuals who are completely informed about the functioning of the monitoring scheme. They argue that the relatively high reward in the case of a positive evaluation drives most of this effect. This is because a positive evaluation at the first interview delays the subsequent monitoring interview by as much as 16 months. Cockx et al. (2011) label this front-loading of job search effort, because the enhanced job search effort induced by the reward must be traded-off against the reduced effort after the interview. The fact that the RDD point estimates are much higher than those estimated within a structural search model can be explained by their imprecision, but can also follow from an incomplete understanding of the monitoring procedure upon receipt of the written notification. Since the monitoring is targeted at long-term unemployed, a non-negligible share of the notified individuals may not be sufficiently literate to fully understand that non-compliance is sanctioned only after two negative evaluations and not earlier than one year after receipt of the notification. The fierce criticism in the media in 2004 by trade-unions and other pressure groups that the new scheme would “hunt down” the unemployed may have had a larger impact on the behavior of this group. This could indeed have enhanced the perception that monitoring was very strict and the threat of a sanction very high, which may therefore have triggered an overreaction of these individuals. The notification also enhanced the transition into training, but not so much as the transition into employment. According to the benchmark model the participation rate increased by 6.5 pp, but for other specifications the effect was found to be lower and not significantly different from zero. This is because many of the unemployed were relatively recently counseled at the regional Public Employment Office, so that if training is more appropriate than job search, it would have started earlier, before the(potential) notification. Nevertheless, we found evidence that the enhanced training participation induced by monitoring could have been a pathway to employment for about one third of the increase in the job finding rate: If indirect transitions to employment via training (or inactivity) are excluded, the treatment effect after eight months falls from 8.8 pp to 6.1 pp. It is unlikely that the unemployed would, in response to the job search requirements, withdraw from the labor force before the moment at which they risk a benefit sanction. Since this moment occurs no earlier than four months after the first evaluation of job search effort, the finding that impact on the transition to inactivity is close to zero and insignificant is consistent with this hypothesis. The finding that monitoring without an immediate risk of a sanction can enhance the job finding rate substantially without stimulating the withdrawal from the labor force is important, since it demonstrates that this outcome can be realized without an immediate income loss for benefit recipients. However, this positive outcome comes at the expense of social costs that are realized ex-post and not analyzed in this research. Some benefit claimants will be sanctioned if they persist in non-complying. Moreover, the front-loading of job search effort prior to the interview comes at an expense of a decline in job search effort ex-post, in the case of a positive evaluation. Whether it is socially optimal to design monitoring schemes with such features depends crucially on the individual and social discount rate and remains to be studied. Finally, our explanation for the magnitude of the treatment effect suggests that it is more the perceived than the effective strictness that determines whether monitoring works. Reaching a better understanding of how these perceptions are formed is therefore an interesting research avenue.